Genome-wide association study shows developmental robustness control by intestinal maltase via internal environment in Drosophila

  1. Soshiro Kashio
  2. Izumi Enoki
  3. Yuto Yoshinari
  4. Takashi Nishimura
  5. Masayuki Miura

Reviewed by

Read the full article See related articles

Discuss this preprint

Start a discussion What are Sciety discussions?

Listed in

Log in to save this article

Abstract

Organisms encounter disturbances during development because of genetic variations, environmental shifts, and stochastic noise. However, developmental robustness and canalization buffer these fluctuations to maintain normal development. Nevertheless, understanding the underlying mechanisms that govern this robustness is challenging because of the complex interactions between these factors. In Drosophila , the number of scutellar sensory organs (SSO; macrochaetes) derived from the sensory organ precursors (SOPs) in the larval wing disc has been used as a model to study developmental robustness. Although the number of SOPs is strictly regulated by a network of signaling and transcription factors in the imaginal disc, non-intrinsic broader factors such as temperature and energy metabolism additionally influence SSO-number fluctuation. Moreover, the precise molecular mechanisms regulating the systemic control of bristle number remain unknown. In this study, we identify factors controlling bristle robustness by performing genome-wide association studies (GWAS). We observed significant single-nucleotide polymorphisms (SNPs) in the Maltase gene cluster and found that the knockdown of Maltase genes affected SSO numbers. Furthermore, Maltase-A1(Mal-A1) in the gut regulated insulin signaling systemically, thereby affecting SSO-number fluctuation. These results suggest that Mal-A1 contributes to robustness by modulating glucose availability and Drosophila insulin-like peptide 3 (dilp3) level, which affects the SOPs in a nonautonomous manner. This study presents the molecular basis of nutritional regulation of developmental robustness and highlights Maltase as a key mediator.

Article activity feed

  1. Review Commons

    Note: This response was posted by the corresponding author to Review Commons. The content has not been altered except for formatting.

    Learn more at Review Commons

    Reply to the reviewers

    The authors do not wish to provide a response at this time.

  2. Review Commons

    Note: This preprint has been reviewed by subject experts for Review Commons. Content has not been altered except for formatting.

    Learn more at Review Commons

    Referee #3

    Evidence, reproducibility and clarity

    The study has carefully controlled and rigorous data. For the most part, the results are consistent with their claims. Except for a few modifications, it should be published. My suggestions are:

    1. Fig 2A. I cannot see the red line in the plot that is mentioned in the legend. Please add it.
    2. Fig 2A. The Manhattan plot shows a number of loci in the genome that have peaks of significant SNPs, not just the locus encompassing Malt-A. It might be worth highlighting the loci or peaks better in the plot. It is pretty minimalist as is.
    3. Linkage disequilibrium is a problem in Drosophila. Many SNPs are hitchikers riding along with a single causative SNP due to infrequent recombination between hitchiker and causative SNPs. How many SNPs are significant and please list the SNPs or intervals considered significant in the GWAS. The text is vague and brief. The plot in Fig 2A is problematic by being overly minimal.
    4. Regarding the GWAS loci they found. It would be worth comparing these regions of the genome with significant GWAS scores to those regions identified in an earlier study. In 2013, Cassidy et al performed artificial selection on Drosophila populations using the same trait (scutellar bristle number) as this study. They did whole genome sequencing of the population before and after selection, and found loci in the genome that exhibited signs of selection through having altered allele frequencies at some loci. Are some of the loci identified in that study the same as in this GWAS study? Are some of the genes implicated in that study the same? The old data is publicly available and so could be easily mined.
    5. Tables 1 is cut apart in its format. Please format properly.
    6. Across the work, there is a lack of statistical testing of significance in bristle number between treated groups. These phenotypes need testing. The number of animals assayed in each experiment are listed but no tests for statistical significance are presented. A chi square or better yet, a fishers exact test would be appropriate. Some of the sample numbers seem low for the claims made, i.e. 8 animals scored for UAS-MalA1 control group.. This testing should be done for all data in Table 1, Fig 2C, Supp Fig 2 A, Fig 4E and any others I might have missed.
    7. Fig 3A, are the individual datapoints single replicates of metabolomic samples? The description of what PCA was done is minimal and needs more description. I assume they performed PCA using metabolites as variables. They did not say. Nor did they explain how PCA was performed except for the software. They "normalized" the data to the median. Did they center the matrix of variable values to the median before doing PCA - is that what they mean? Why not center to the mean values? Typically one calculates the mean value for a given variable, ie a single metabolite, across all samples, and then calculates the difference between the measured value from one sample and the mean value for that variable. That needs to be done. It is not standard to center to the median. They should also normalize the data to eliminate biasing in the PCA results because of variance due to very abundant metabolites, The variables with large values (ie abundant metabolites) overly contribute to the explanatory variance in a PCA analysis unless one normalizes. This normalization is typically done by taking the difference between measured and mean values (as described above), and dividing that difference by the standard deviation of the variable's measurements. Think of it as a Z-score. The matrix data then is centered around zero for each variable, and each variable's values range from -5 to +5. Then perform PCA. Otherwise highly abundant metabolites bias the analysis. Again, this type of normalization is standard for PCA.
    8. How many metabolites were measured? What were they, ie the list. Provide please
    9. Results described in Fig 5A are the weakest in the manuscript and really could be supplemental. It is weakly circumstantal evidence for the claim being made. Temperature affects so many things, it could be coincidence that dilp levels change and this change correlates with bristle number. Many things change with temperature. Definitely they should not end the results section with such weak data,
    10. Carthew and colleagues showed that IPC ablation suppressed the scutellar bristle phenotypes of miR9a and scute mutants. Does Mal-A1 knockdown have similar effects on these mutants? One would predict yes.
    11. The authors mention the 2019 paper by Cassidy et al and some of the results therein regarding inhibiting carbohydrate metabolism and phenotype suppression (robustness). But not only miR-9a and scutellar bristles were tested in that paper but a wide variety of mutations in TFs, signaling proteins and other miRNAs. All their results were consistent with the findings of the current ms. The authors could discuss this more in depth. Also, Cassidy et al put forth a quantitative model that explained how limiting glucose metabolsm could provide robustness for a wide variety of developmental decisions. It might be worth discussing this model in light of their results.

    Significance

    This manuscript describes an interesting study of developmental robustness and its intersection with organismal metabolism. It builds upon prior papers that have addressed the link between metabolism and development. It describes an ingenious approach to the problem and uncovers maltose metabolism in Drosophila as one such connection to sensory organ development and patterning. The important take home message for me is that they found natural genetic variants from the wild that confer greater robustness to the fly's morphological development, and these genetic variants are found in an enzyme that broadly metabolizes maltose, a simple sugar. Whereas previous studies used genetic manipulation to impact metabolism, this study shows that genetic variants in the wild exhibit effects on robustness. It suggests there might be a tradeoff between more vigorous carbohydrate metabolism and fidelity in morphological development.

  3. Review Commons

    Note: This preprint has been reviewed by subject experts for Review Commons. Content has not been altered except for formatting.

    Learn more at Review Commons

    Referee #2

    Evidence, reproducibility and clarity

    Summary:

    In this study, the authors performed GWAS to identify associations between the mean bristle number in Drosophila melanogaster adults and different SNPs present in 95 lines of the DGRP panel rear at 18C. They selected genes harboring those SNPs linked to bristle number that also had a moderate or high expression at the third insta larva stage to perform an RNAi screen. This RNAi screen, which included 43 genes, identified Maltase-A1 (Mal-A1) as a contributor to bristle number. Therefore, the authors then focus on investigating possible metabolic and transcriptional changes underlying the effect of Mal-A1 knockdown on bristle number. After whole-body knockdown using the da-gal4 driver, the authors identified decreased glucose in whole body and hemolymph, and decreased dilp3 mRNA expression in whole body, intestine, and insulin producing cells (IPC) in the larva brain. Similar to a whole-body Mal-A1 knockdown, a gut epithelial cell-specific gal4 driver (NP1) also decreased dilp3 mRNA expression in the whole body and larva brain. The authors suggest that Mal-A1 activity in the intestine may affect bristle number through lowering available glucose in the intestine, which decreases circulating glucose levels in the hemolymph, and in turn decreases dilp3 mRNA expression in the larva brain, leading to decreased bristle number. Finally, to validate the influence of bristle number via dilp3-mediated insulin signaling in the brain, the authors reared larvae at 18C, which they showed increased bristle number. Supporting their proposed model, rearing larvae at 18C increased dilp3 mRNA expression in the brain, which correlated with increased bristle number.

    Major comments:

    1. The main finding of this paper is the identification of Mal-A1 gene as a regulator of bristle number in Drosophila adults. However, the authors do not to show clear phenotypes which could stem from a lack of experimental rigor. As an example in Fig. 2C (source data not provided) the UAS-Mal-A1-RNAi line V15789 in the absence of GAL4 shows 5% abnormal bristle number compared with 2% upon knockdown. If I'm understanding the data provided, this means that abnormal bristle number was observed in 2 flies (out of 40) in the UAS-line alone compared with ~2 flies (out of 111) in the presence of GAL4. For line V106220, 2% (n=56) showed abnormal bristles compared with 0% (n=37) upon in the presence of GAL4. In absolute numbers this would mean that abnormal bristle number was observed in ~1 fly (out of 56) in the UAS-line alone compared with 0 flies (out of 37) upon knockdown. All of these experiments do not use sufficient n, which according to the reviewers calculations (to show a 3% increase, with 80% confidence the n should be around 750-800). In addition no information on statistical tests or whether biological replicates were performed is included. Due to the main finding heavily relying on this phenotype of abnormal bristle number, this reviewer is not confident that the conclusions of the manuscript are supported. This problem also applies to other experiments presented in the manuscript, which suffer from low n, significantly decreasing the enthusiasm for the presented results.
    2. The authors do not to show that Drosophila insulin- like peptide 3 (dilp3) level affects the SOPs in a nonautonomous manner. The only experiments included are showing indirect effects.
    3. There are important statistical details missing in some of the figures (see comments below)
    4. Important details are missing from the methods for results or analysis to be reproduced. For example, the method section for GWAS analysis is lacking details, a script should be provided as supplemental information, as well as a table similar to the one provided for the RNAi screen.

    Minor comments

    • There are some typos like referring to 'using w118 male mice' in the 'Phenotypic Analysis of Maltase Knockdown; (1) Bristle number count'
    • Details in methods. For GWAS experiments, could the authors define what their cutoffs were for selecting genes harboring SNPs linked to bristle number? How many base pairs from a gene? or enhancer? They selected only those gene with moderate or high expression, but what does it mean?
    • In Fig. 2A, could the authors provide all significant SNPs identified by their GWAS analysis as supplemental material?
    • In Fig. 2A, it is stated in the legend " and the red line represents the significance threshold calculated using Bonferroni correction...". This might be a problem with the pdf document but I did not find the red line in the Manhattan plot that the authors refer to.
    • In Fig. 4E, could the authors provide the n number as in other figures?
    • Check citations. Some references have missing parts. For example; Ref 5 is missing the last 2 words of the title. In Manuscript it reads: "Trehalose metabolism confers developmental robustness and stability in Drosophila by regulating.". It should be "Trehalose metabolism confers developmental robustness and stability in Drosophila by regulating glucose homeostasis."

    Significance

    While the significance of identifying a novel regulatory mechanism for developmental robustness in Drosophila melanogaster is high and would be interesting for a broad audience, the authors do not present convincing experimental evidence to support their hypothesis. This is due to the insufficient number of replicates as well as the lack of experiments showing a direct role of insulin signaling.

  4. Review Commons

    Note: This preprint has been reviewed by subject experts for Review Commons. Content has not been altered except for formatting.

    Learn more at Review Commons

    Referee #1

    Evidence, reproducibility and clarity

    In this article, the authors identified a gut-expressed enzyme, maltase-A1, which regulates the developmental robustness of SSO. Through a GWAS analysis of bristle phenotypes using the DGRP lines, maltase-A1 was revealed as a significant regulator of bristle number. Knockdown of maltase-A1 suppressed the increase in bristle count. Metabolite profiling of key carbohydrates showed a reduction in several sugar levels, potentially leading to decreased ilp3 release from the CNS. Furthermore, the authors demonstrated that cold temperature induces higher expression of both ilp2 and ilp3, which may contribute to the observed increase in bristle number.

    The findings of this study offer valuable insights into how nutrient metabolism may influence developmental robustness. However, a key limitation lies in the correlative nature of the observations. Further experimental validation is needed to establish a direct causal relationship.

    1. To further support the hypothesis that Maltase-A1 mutation reduces bristle variance by lowering hemolymph glucose levels and consequently decreasing insulin secretion, it would be essential to test whether providing a higher concentration of dietary glucose to Maltase-A1 mutant larvae can rescue the mutant phenotype.
    2. To further substantiate the claim that ilp3 acts as a downstream effector in the Maltase-A1 regulatory pathway, it would be important to perform ilp3 knockdown and overexpression experiments in both wild-type and Maltase-A1 mutant backgrounds. This approach could help determine whether altered ilp3 expression levels directly contribute to the bristle phenotype associated with Maltase-A1 dysfunction.

    Significance

    How nutrients regulate developmental processes is an intriguing question in developmental biology. This study employs GWAS to identify an unexpected regulator of bristle development, offering new insights into how nutrient metabolism may influence developmental robustness. I believe this article will be of great interest to audiences in developmental biology.

  5. Version published to 10.1101/2025.08.13.669836 on bioRxiv